EXCEED THE SPACE PROVIDED. Postmortem and in vivo MRI studies of autism have identified multiple loci of anatomic abnormality, including the cerebellum, frontal cortex and parietal cortex. Research has also uncovered multiple domains of neurobehavioral abnormality, including motor, sensory, attention, learning, memory, language, social and affective. Attempts to explain the relationship between anatomic abnormalities and functional deficits have led to numerous hypotheses. However, many of these have been founded on inference from one type of evidence (e.g., behavioral) in the absence of concomitant evidence from other paradigms (neurofunctional and neuroanatomical). In view of the scarcity of brain-behavior experiments directly correlating neurobehavioral with structural and functional brain measures, hypotheses to date remain largely speculative. In a series of recent brain-behavior experiments which compare anatomic and cognitive-behavioral measures of the same patients, we have begun to elucidate the relationship between specific anatomic defects and specific functional deficits in autism. Thus, we have demonstrated that cerebellar anatomic abnormality is associated with deficits in visual selective attention, motor control, sensorimotor learning, visuospatial exploration, and orienting attention. FMRI activation patterns suggest deviant organization of motor and attention functions in the autistic cerebellum. We have further demonstrated the correlation of parietal anatomic abnormality with abnormalities in selective tuning of visuospatial attention. Also, ERP maps suggest aberrant frontal and temporal topographic distributions of spatial attention-related components. Our brain-behavior evidence is consistent with the general hypothesis that autism involves aberrant functional organization in cerebellar, frontal, parietal and temporal cortices, and these defects underlie multiple cognitive-behavioral deficits. In our proposed brain-behavior experiments, therefore, we will elucidate the brain bases of autism by correlating measures of anatomic abnormality (MRI) with measures of functional abnormality (fMRI, ERP) and deficits in various cognitive domains, including attention, learning, memory, language and 'executive' functions. In this way, we will test brain-behavior links not previously examined in autism, as well as those that have only recently been reported. Further, we will compare the patterns of functional activation in cerebral and cerebellar cortices in autistic versus normal subjects during the performance of tasks involving these different domains. We hypothesize that cerebellar and cerebral cortices will show abnormal fragmentation of functional specialization in autism.